.DOCX File Format – The package

Parts and part names

A part name is the member’s path with a leading /, such as /word/document.xml. Part names are compared without regard to the case of ASCII letters, and may contain percent-encoded characters. ZIP folders have no meaning in a package: only the parts do.

Word keeps a conventional layout, but other writers may use other names, so a reader finds the parts through the relationships and not by name. The usual parts of a document are:

[Content_Types].xml

This part, at the root of the archive, gives the media type of every part. A Default element gives the type for an extension. An Override gives the type of one part and takes priority over a Default.

<Types xmlns="http://schemas.openxmlformats.org/package/2006/content-types">
  <Default Extension="rels" ContentType="application/vnd.openxmlformats-package.relationships+xml"/>
  <Default Extension="xml" ContentType="application/xml"/>
  <Default Extension="png" ContentType="image/png"/>
  <Override PartName="/word/document.xml"
    ContentType="application/vnd.openxmlformats-officedocument.wordprocessingml.document.main+xml"/>
  <Override PartName="/word/styles.xml"
    ContentType="application/vnd.openxmlformats-officedocument.wordprocessingml.styles+xml"/>
</Types>

The main part’s content type tells the four kinds of Word file apart:

Other WordprocessingML parts have types of the form application/vnd.openxmlformats-officedocument.wordprocessingml.styles+xml: styles, settings, webSettings, fontTable, numbering, footnotes, endnotes, comments, header and footer. The glossary document is …wordprocessingml.document.glossary+xml.

Relationships

A part’s relationships are in a part beside it: the relationships of /word/document.xml are in /word/_rels/document.xml.rels. Those of the package itself are in /_rels/.rels.

<Relationships xmlns="http://schemas.openxmlformats.org/package/2006/relationships">
  <Relationship Id="rId1"
    Type="http://schemas.openxmlformats.org/officeDocument/2006/relationships/styles"
    Target="styles.xml"/>
  <Relationship Id="rId2"
    Type="http://schemas.openxmlformats.org/officeDocument/2006/relationships/hyperlink"
    Target="https://www.example.com/" TargetMode="External"/>
</Relationships>

The relationship types used in Word files, all with the prefix http://schemas.openxmlformats.org/officeDocument/2006/relationships/ unless shown otherwise:

Because the target of a relationship can be any URI, a document can point outside itself: a linked picture, a remote template. The viewer on this site lists such targets and never fetches them.

Reading a package

1. Read the ZIP central directory and find [Content_Types].xml.
2. Read _rels/.rels and follow the officeDocument relationship to the main part.
3. Check that part’s content type, and that its root element is w:document.
4. Read the main part’s relationships for its styles, numbering, headers, pictures and so on, and follow only those it uses.

Members of a package are normally Deflate-compressed or stored, and the package format does not encrypt anything. A password-protected Word file is not a ZIP archive at all: it is a Compound File that holds the encrypted package in a stream named EncryptedPackage.

The ZIP structure itself is described on the ZIP archive pages.